Exploring the Rapid Assessment Method for Nature Reserve Landscape Protection Effectiveness—A Case Study of Liancheng National Nature Reserve, Gansu, China

The rapid assessment of the effectiveness of landscape protection in nature reserves is of great significance for the scientific formulation of protection and management countermeasures. There is also an urgent problem to be solved for the construction and management of nature reserves in China. Using high-resolution remote sensing image data, this study analyzes the landscape dynamics in the Liancheng National Nature Reserve (LNNR) and their driving factors since the reserve’s promotion to the national level in 2005, and proposes a comprehensive evaluation method for the effectiveness of landscape protection in protected areas based on the Landscape Transfer Index (LTI), Protected Landscape Integrity Index (PLII), and Interfered Landscape Sprawl Index (ILSI). Between 2006 and 2019, the area of protected landscape—namely woodland, grassland, and water—in the LNNR decreased, while the area of interfered landscape such as residential land, industrial and mining land, and water conservancy facility land increased. The LTI was −0.14, and among the driving factors, the development of industry and mining, land use by indigenous inhabitants, and the development of the transport industry made the highest contributions to the landscape transfer tendency, respectively 34.79%, 28.98%, and 17.30%. The results indicate that the overall quality of the landscape declined slightly during this period, mainly as a result of industrial and mining development, indigenous use of the land, and road construction. The PLII decreased from 82.7 to 68.7; this result indicates that the spatial pattern of the protected landscape became more fragmented, and the degree of human interference in the landscape increased. The ILSI increased from 26.61 to 26.68 which indicates that the scope of influence of human interference did not change significantly. The effectiveness of landscape protection in the LNNR is low. Despite the insignificant nature of these changes, they still require attention and timely remedial measures. The methodology proposed in this study may be applicable to the rapid assessment of the effectiveness of landscape conservation in various types of nature conservation sites around the world.


Introduction
Natural habitat loss and fragmentation are important threats to global biodiversity conservation [1,2]. Studies have shown that the establishment of protected areas (PAs) can effectively curb the destruction of natural ecosystems, and this approach is a major component of biodiversity conservation [3][4][5]. In 2019, there were 261,200 PAs of various types in the world, covering 15.3% of the global terrestrial and freshwater environment [6]. Analyzing landscape changes in PAs is an intuitive way to monitor and evaluate such areas. Scholars classify the landscapes in PAs into two categories according to their attributes, namely protected and interfered landscape [7][8][9]. The former mainly refers to natural landscapes but also includes some human-made landscapes that are beneficial to wildlife; the latter refers to the types of human-made landscape that causes interference and damage to natural ecosystems and wildlife habitats [10][11][12]. It is generally accepted that PAs are effective for the conservation of natural habitats [13,14]. However, some scholars have argued that the loss of wildlife habitat in some PAs due to increased anthropogenic interference has still not been effectively curtailed, questioning the conservation effectiveness of nature reserves [15,16]. Therefore, it is highly necessary to scientifically assess the conservation effectiveness of PAs by monitoring the dynamic changes in the protected landscape.
In China, there are 11,800 PAs of various types, covering a total area of more than 1.8 million km 2 . A system of nature reserves in China consisting of national parks (pilot), nature reserves, nature parks (including scenic spots, forest parks, geoparks, natural cultural heritage, wetland parks, desert parks, aquatic germplasm reserves, marine parks, marine special reserves, nature conservation plots, etc.) has been formed [17]. Among them, nature reserves occupy the largest area among the various types of PAs and are of fundamental importance for ecological conservation [18]. Since the establishment of the Dinghushan Nature Reserve in 1956, there are a total of 2750 nature reserves in China, covering a combined area of 1.47 million km 2 , accounting for 15% of China's land area; of these, 474 are state-level nature reserves covering a total area of 980,000 km 2 , accounting for 10% of China's land area [19]. As of 2015, anthropogenic activities are prevalent across 446 national reserves in China [20], with expectations of continued increase as the population increases [21]. Furthermore, there is a lack of quantitative assessment of the effectiveness of landscape protection within China's nature reserves, and how to scientifically and rapidly assess the effectiveness of landscape protection within nature reserves following their establishment is an urgent problem for the construction and management of nature reserves. Research by Chinese and foreign scholars on the conservation effectiveness of nature reserves at the landscape level has mainly focused on changes in landscape dynamics, the impacts of human interference, and the effectiveness of habitat quality protection [14,[22][23][24][25][26]. These studies evaluated the conservation effectiveness of different types of nature reserves from different perspectives; however, a fast, effective, and universal method to evaluate the landscape protection effectiveness of nature reserves is lacking.
In recent years, scholars in China and abroad have conducted a large number of studies on the dynamic changes of regional landscapes and their driving factors using a combination of remote sensing and GIS, based on theories related to landscape ecology [27][28][29][30]. Most of these studies were conducted at large spatial scales using moderate-resolution (≥30 m) satellite image data. In nature reserves, under strict control measures of the Regulations of the People's Republic of China on nature reserves (http://www.forestry.gov.cn/main/3950/content-459882.html, accessed 15 January 2021), the landscape generally does not change to a large extent; however, smaller-scale anthropogenic activities that lead to landscape changes are common, such as the construction of water conservancy facilities, residential land use, and tourism service facilities. It is difficult to identify these activities with medium-resolution satellite imagery. With the development of remote sensing and geographic information systems, high-resolution satellite imagery has enabled researchers to accurately identify landscape changes in nature reserves at smaller scales [31][32][33], providing the possibility to accurately assess the effectiveness of landscape protection in nature reserves.
Beijing Forestry University is developing the Technical guidelines for assessing conservation efficiency of nature reserves, which aims to provide a rapid and effective evaluation method for assessing the effectiveness of nature reserves. In the section on landscape protected effectiveness, a method based on the Landscape Transfer Index (LTI), the Protected Landscape Integrity Index (PLII), and the Interfered Landscape Sprawl Index (ILSI) is proposed to quickly evaluate the effectiveness of landscape protection. This study takes the Liancheng National Nature Reserve (LNNR) as the study area and uses high-resolution remote sensing image data from 2006 and 2019 to analyze the landscape change trend and its drivers from the perspective of landscape type change since the nature reserve was promoted to a national nature reserve in 2005; moreover, it proposes the use of the PLII and the ILSI to attempt to quantify the integrity of protected landscape and the influence range of human disturbance from the perspective of the landscape spatial distribution pattern. These methods were used to comprehensively evaluate the effectiveness of landscape protection in the LNNR since its promotion to a national nature reserve. The main objectives of this study are as follows: (1) To propose a method to rapidly assess the effectiveness of landscape protection in nature reserves; (2) to evaluate the effectiveness of landscape protection in the LNNR using this method; and (3) to propose countermeasures and recommendations for landscape protection management in the LNNR.

Study Area
The LNNR is located in the arid northwest of China, in Yongdeng County, Lanzhou City, and in the Tianzhu Tibetan Autonomous County, Wuwei City ( Figure 1 (42.43%) is the experimental zone. The LNNR is located in the middle and lower reaches of the Huangshui River, a major tributary of the Yellow River and has been summarized as "A river in the middle of two mountains" [34] as the Datong River divides the whole area into two parts, East and West, with different geomorphic types. The LNNR is also located in the transitional area between the East extension of Qilian Mountains and the West Gansu subsidence basin, which is characterized by the staggered distribution of steep rocky mountains from Qilian Mountain in the West and loess mountains from the Loess Plateau in the East. The altitude gradually increases from east to west, showing the topographic characteristics of high altitude in the West and low in the East. The LNNR has characteristics of temperate continental monsoon climate with an average annual precipitation of 419 cm and is a forest ecosystem type nature reserve with large areas of primeval forest composed of S. przewalskii, Picea wilsonii, and P. crassifolia. The large focus of protection of its Sabina przewalskii and Picea crassifolia forest ecosystem and the habitat of nationally protected wild animals and plants makes LNNR an important forest and water conservation area in the upper reaches of the Yellow River. For example, the LNNR has a wide variety of vegetation, which can be divided into six vegetation types, 17 formation groups, and 28 formations, providing a good habitat for wild animals and plants to survive and reproduce [34]. Therefore, the LNNR is not only regionally representative but also has high conservation and scientific research value. As of 2019, there were 30,695 community residents living within the LNNR, distributed among 10 administrative villages and one community in two townships, mainly in the experimental zone of the reserve. These communities are subject to frequent human activities, which greatly threaten the protection and management work of the reserve.

Data Sources
The remote sensing images used in this study were obtained in August 2006 and May 2019, respectively. The images in 2006 were taken by SPOT-5 satellite that was successfully launched in 2002, these images include 2.5-m panchromatic band and 10-m multispectral bands. The multispectral band images were fused from 10 m to 2.5 m in resolution based on the panchromatic band image using Gram-Schmidt Pan Sharpening tool in ENVI. The high-resolution cloud-free images (2-m resolution) in 2019 was taken by China's Gaofen-1 (GF-1) satellite that was successfully launched in 2013.

Remote Sensing Image Processing and Landscape Classification
Using the ENVI 5.1 software (L3Harris Geospatial, Boulder, Colorado, USA) and the ArcGIS 10.2 software, the two image series were processed with geometric correction, multiband blending, panchromatic sharpening, and cropping. The protected area was divided into 10 primary landscape types and 21 secondary landscape types (Table 1) based on the Chinese Land Use Classification Standard (GB/T 21010-2017). Using an example-based object-oriented classification method to classify the 10 primary landscape types by ENVI 5.1 software. The specific steps are as follows: First, the Edge-Based Segmentation algorithm was used, i.e., after repeated experiments the threshold of segmentation was set to 45, and the better edge characteristics could be obtained by segmentation. Second, the Full Lambda-Schedule algorithm was used, the threshold of mergence was set to 90 and then adjacent small patches were iteratively merged based on spectral and spatial information. Third, the attributes of image space, spectrum, texture, color space, and band ratio were calculated. Fourth, the classification samples of the 10 primary landscape types were selected based on the field data collected in July 2018. Fifth, the K Nearest Neighbor method was used in the classification algorithm and images classified considering the adjacent five pixels. Sixth, the clustering and filtering method was used to remove background noise. Any regional pixels wrongly classified or missed were manually corrected based on the field survey data. Seventh, the confusion matrix was computed using the validation samples selected from the SPOT-5 fusion image and GF-1 HR image, and the classification accuracy was tested. Manual visual interpretation (i.e., direct delineation and interpretation of the images) in ArcGIS 10.2 was used to classify the 21 secondary landscape types based on remote sensing image data and field survey data. The accuracy of the decomposition results was assessed using a field survey, and the attribute accuracy and boundary accuracy of the decomposition results for 2006 and 2019 were found to be greater than 97% (Table 2), which meets the accuracy requirements. The 21 secondary landscape types were classified into protected, interfered, and neutral landscapes based on the attributes of the land use/land cover (LULC) types (Table 1): protected landscapes include natural forest, artificial forest, shrub forest, grasslands, and river water surface; interfered landscapes include agricultural land, mining land, industrial land, roads, hydraulic architecture, urban residential land, detached houses, rural residential land, and other construction land. Neutral landscapes include bare land and bare gravel land. Based on the ecological function and role in habitat conservation, the natural and artificial landscape was further divided into three functional types: protected landscape, interfered landscape, and neutral landscape. The golden section method was used to classify the ecological level of landscape into seven grades: 1.00, 0.62, 0.38, 0, −0.38, −0.62, and −1.00 (Table 1). The golden section method was a typical algorithm in selecting optimization based on Golden Section Theory [35], and it used the critical values of 0.62 and 0.38 to divide the given interval, in which threshold values of 0.62 and 0.38 were used to divide the given interval. A higher assigned value indicates a higher ecological level. Interfered landscapes have negative values, and larger absolute values indicate a higher level of anthropogenic interference. Neutral landscapes are assigned a value of 0 [7].

Landscape Transfer Matrix
The intersection module of ArcGIS was used to calculate the geometric intersection in 2006 and 2019. Landscape-type vector maps were used to statistically derive the area change of each landscape type between different periods and generate the landscape transfer matrix. To clearly demonstrate the changes in landscape pattern, the networkD3 package for R was used to draw the landscape transfer matrix Sankey diagram, and the area change rate of each landscape type was also calculated.

Landscape Transfer Index
The LTI is an overall characteristic value of the effectiveness of landscape protection, which directly reflects the overall trend of landscape type change and the dynamics of land use type to reflect the level of landscape protection. The LTI of the LNNR was calculated based on the landscape transfer matrix using the following formula: where I T is the landscape transfer index; m is the number of landscape types in the previous period; n is the number of landscape types in the current period; S i→j is the area in which landscape type i was converted to landscape type j; S T is the total area of the nature reserve; and D i and D j denote the assigned ecological levels of landscape types i and j, respectively. The LTI has values between −200 and 200; positive values indicate that the landscape and ecological conditions of the area have improved and negative values indicate that they have deteriorated, with larger absolute values indicating greater change and vice versa.

Protected Landscape Integrity Index
In this study, the PLII is used to characterize the integrity of the habitats of the main protected objects in nature reserves, which are a mosaic of protected landscapes such as woodlands, grasslands, and waters. The fragmentation index and the edge effect index were used to comprehensively characterize the degree of integrity of the protected landscape in the LNNR using the following formulas: where I I is the integrity index of the protected landscape; S T is the total area of the LNNR; I F is the fragmentation index of the protected landscape, reflecting the overall fragmentation of all protected landscapes, and has values of between 0 and 1; S j is the area of the jth protected landscape mosaic patch; n is the total number of protected landscape mosaic patches; I E is the edge effect index of the protected landscape; P j is the perimeter of the jth protected landscape mosaic patch.
The value of I I ranges from 0 to 100, with larger values indicating less fragmentation of the protected landscape mosaic, weaker edge-effect strength, and greater landscape integrity, and vice versa.

Interfered Landscape Sprawl Index
The ILSI is used to characterize the spreading degree and edge effect intensity of land use types, such as Cultivated land, Residential land, Industry and storage land, and Water conservancy facility land, that disturb and destroy the natural landscape in nature reserves, and to evaluate the impact range of anthropogenic interference in the nature reserves. This index was calculated according to the following formulas: where I S is the interfered landscape sprawl index; P ij is the ratio of the common boundary length (E ij ) between the i-th interfered landscape mosaic patch and the j-th mosaic patch for all other types of landscape (including protected and neutral landscapes) to the total common boundary length (E) between all interfered landscape mosaic patches and other landscape mosaic patches; m is the total number of interfered landscape mosaic patches; and n is the total number of mosaic patches for all other types of landscape. The value of I S ranges from 0 to 100; the larger the value, the more dispersed the distribution of interfered landscape, the greater the edge effect, and the greater the extent of anthropogenic interference.

Classification of Drivers of Landscape Change
To analyze the causes of landscape changes in the LNNR, the landscape change patches in different periods were extracted using the ArcGIS topological analysis module. According to the actual situation of the LNNR obtained through past field investigation and image processing, the direct driving factors of each landscape patch change were qualitatively classified and the direct driving factors leading to landscape change were classified into nine categories (Table 3).

Changes in Landscape Area
The LNNR has three main types of protected landscapes (forest land, grassland, and water) and six types of interfered land (cropland, residential land, traffic facility land, water conservancy facility land, industrial and mining land, and other construction land). The results show that, between 2006 and 2019, the total area of protected landscapes in the reserve decreased from 42,505.59 hm 2 (88.25% of the reserve's total area) to 42,456.16 hm 2 (88.15%) ( Table 4). Woodland occupies the largest area among all landscape types in the LNNR, and its area decreased from 41,325.16 hm 2 (85.80% of the reserve's total area) in 2006 to 41,292.32 hm 2 (85.73%) in 2019, a decrease of 32.84 hm 2 . In the same period, the grassland area decreased from 771.03 hm 2 (1.6% of the reserve's total area) to 760.86 hm 2 (1.58%), while the water area decreased from 409.4 hm 2 (0.85% of the reserve's total area) to 402.98 hm 2 (0.84%). That is, between 2006 and 2019, the areas of all types of protected landscapes in the reserve decreased slightly, and the annual rate of change in area was small (Table 4). In the same period, the area of interfered landscape in the reserve increased from 5526.36 hm 2 (11.47% of the reserve's total area) to 5546.71 hm 2 (11.52%). Of the various types of interfered landscape in the LNNR, cropland accounts for the largest area, and decreased from 4808.01 hm 2 (9.98% of the reserve's total area) in 2006 to 4773.51 hm 2 (9.91%) in 2019. During the same period, all other types of interfered landscape also increased in area to varying degrees, with residential landscape experiencing the largest percentage of area expansion of any type of interfered land, increasing continuously from 462.38 hm 2 in 2006 to 488.45 hm 2 in 2019. Water facility landscape experienced the fastest area expansion of any interfered landscape type, with a mean annual change in area of 5.54%.

Area Transfer between Different LANDSCAPE Types
Between 2006 and 2019, the area of protected landscape types such as natural forest, shrub forest, grassland, and river water surface in the LNNR decreased. As shown in the landscape transfer matrix (Table 5) and the landscape transfer Sankey diagram (Figure 2), natural forest was mainly converted to bare land and bare gravel land with 11.41 hm 2 and 6.86 hm 2 , corresponding to proportions of 47.96% and 28.84%, respectively. Shrub forest was mainly converted to bare land and detached houses, with conversion areas of 4.40 hm 2 and 1.99 hm 2 , corresponding to proportions of 34.24% and 15.49%, respectively. Grasslands were mainly converted to cultivated land and ditches, with conversion areas of 4.29 hm 2 and 2.19 hm 2 , corresponding to proportions of 28.32% and 14.46%, respectively. River water surface was mainly converted to hydraulic architecture, mining land, and highways, with conversion areas of 1.49 hm 2 , 3.72 hm 2 , and 2.69 hm 2 , corresponding to proportions of 12.58%, 31.42%, and 22.72%, respectively.
Additionally, between 2006 and 2019, the area of cultivated land in the interfered landscape in the LNNR decreased significantly. This cultivated land was mainly converted to bare land and rural residential land, with conversion areas of 16.84 hm 2 and 13.7 hm 2 and conversion rates of 30.63% and 24.92%, respectively. Furthermore, the areas of other types of interfered landscape, such as industry and storage land, hydraulic architecture, and highways, all increased to varying degrees. The industry and storage land in the protected area are ore processing workshops or ore dumps, which have mainly been converted from bare land, with a conversion area of 3.11 hm 2 and a proportion of 63.34%. The hydraulic architecture land is mainly used for dams or power stations, and has mainly been converted from river water surface, with a conversion area of 1.49 hm 2 and a transfer ratio of 85.14%. Highways were mainly converted from cultivated land and river water surface, with a conversion area of 3.16 hm 2 and 2.69 hm 2 , accounting for 36.36% and 30.96%, respectively.  Moreover, the areas of the two types of neutral landscape, bare land and bare gravel land, increased significantly between 2006 and 2019, with bare land mainly being converted from natural forest and cultivated land, with conversion areas of 11.41 hm 2 and 16.84 hm 2 , accounting for 31.72% and 46.82%, respectively. The bare gravel land is mostly distributed in the rocky hills on the west bank of the Datong River, and the main source of transfer is from natural forest, with a transferred area of 6.86 hm 2 , corresponding to a conversion rate of 91.34%.
The calculated LTI of the LNNR is −0.14, indicating that the landscape of the reserve tended to deteriorate in general during 2006-2019.

Analysis of Landscape Transformation under Different Drivers
Between 2006 and 2019, the total area of changed landscape patches was 135.44 hm 2 and the total number of changed patches was 291 (Figure 3). Seven of the nine driving factors were found to contribute negatively to the LTI (Figure 4), among which the absolute values of the LTI caused by the development of industry and mining, land use by indigenous inhabitants, and the development of the transportation industry were high, with contributions of 34.79%, 28.98, and 17.30%, respectively. With socioeconomic development and population increase, the demand for construction land increased, and the number of landscape change patches caused by land use by indigenous inhabitants was the largest (135), with a total area of 37.8 hm 2 . Land use by indigenous inhabitants was the largest driving factor of landscape area change and was also one of the main causes of the LTI decrease. In the experimental zone in the south of the reserve, there are large state-owned industrial and mining enterprises with a complete production line of ore mining, stockpiling, and processing built in the 1970s; this has caused some of the natural forest, shrub forest and water in the experimental zone of the reserve to degenerate into bare land or industry and storage land. The development of industry and mining has been the main driving factor for landscape deterioration in the reserve. There are 77 km of roads in the LNNR; the main transportation road in the reserve is provincial highway S301, which passes through the central experimental zone of the reserve, following the Datong River. From 2006 to 2019, 3.8 km of road reconstruction and expansion was carried out in the LNNR, mainly focusing on the provincial highway S301, which transformed part of the shrub forest, water, and cultivated land along the Datong River into transportation land. Additionally, there are four hydropower stations in the reserve, all of which are located in the reserve's experimental zone; of these, two were newly built, and one was expanded, between 2006 and 2019. Water facility land is the landscape type that experienced the fastest change in LNNR during this period. However, the contribution of the development of the hydropower industry to the LTI decrease is insignificant (3.16%). This can be attributed to the fact that the construction of hydropower plants led to the conversion of protected landscapes (e.g., river water surface) into construction land; however, it also increased the water level, and consequently, patches with low assigned ecological levels along rivers (e.g., bare land and arable land) were converted into water lands, which have a higher assigned ecological level.
Between 2006 and 2019, the LTI caused by population displacement and ecological restoration are positive, being 14.61 × 10 −3 and 12.37 × 10 −3 , respectively. The landscape change due to population displacement was mostly a shift from interfered to neutral landscape and did not produce a significant improvement in landscape quality. Although ecological restoration led to a significant improvement in landscape quality, the area of implementation is small. Therefore, the population displacement and the ecological restoration were not sufficient to compensate for the overall trend of poorer landscape quality due to other drivers.

Protected Landscape Integrity
The PLII of the LNNR decreased significantly between 2006 and 2019, from 82.7 to 68.7, indicating that the integrity of the protected landscape of the LNNR decreased significantly during this period.
Between 2006 and 2019, the total area of the protected landscape mosaic in the LNNR decreased by 49.43 hm 2 (0.12%), from 42,505.59 hm 2 to 42,456.16 hm 2 (Table 6); additionally, the number of patches increased from 109 to 128, the fragmentation index increased significantly, from 0.22 to 0.57, and the edge effect index decreased slightly, from 1.18 to 1.15, indicating that the fragmentation of the protected landscape mosaic was the main reason for the reduced integrity of the protected landscape in the LNNR between 2006 and 2019. The main road network of the LNNR comprises the S301 provincial highway, in the central part of the experimental zone of the LNNR, the X137 county road, in the east of the LNNR, and a tourist road in the west of the Tulugou National Forest Park. In 2006, some road sections were narrow and not hardened, and did not form a sufficiently resistant surface, and therefore the protected landscape mosaic was not spatially completely cut away. Between 2006 and 2019, the reconstruction and expansion of roads resulted in a significant increase in the fragmentation of the protected landscape mosaic and a significant decrease in the landscape connectivity.

Interfered Landscape Sprawl
The ILSI of the LNNR was 26.61 and 26.68 in 2006 and 2019, respectively. The small change in the index over the study period indicates that the scope of influence of interfered landscape in the LNNR did not change significantly.
Between 2006 and 2019, the total area of interfered landscape in the LNNR increased by 20.35 hm 2 (0.05%), from 5526.36 hm 2 to 5546.71 hm 2 , with an area conversion rate of 0.37%; additionally, the number of interfered landscape mosaic patches increased from 116 to 131 (Table 7). The degree of dispersion of the spatial distribution of the interfered landscape changed very little, and the increase in area occurred mostly in areas where there was originally a high concentration of interfered landscape; furthermore, the new patches were mostly small, which suggests that the control of the scope of influence of anthropogenic interference in the protected area of the LNNR was effective.

Discussion
From 2006 to 2019, the area of protected landscape in the LNNR decreased and the area of interfered landscape increased; that is, the landscape in the reserve generally tended to deteriorate and the conservation effectiveness of the reserve was not significant. The implementation of major national natural forest resource protection projects promotes the development of natural ecosystems. In the LNNR, most of the protection project tasks, such as the natural forest protection projects, were completed before the reserve was promoted to a national nature reserve in 2005; natural forest logging has been forbidden since 1998. Additionally, 446 hm 2 of sparse forest land and shrubbery land in the reserve were afforested between 1998 and 2005. From 2001 to 2005, a reforestation project of 703.5 hm 2 was completed in the eastern experimental zone of the reserve. Unfortunately, we do not have access to high-resolution remote sensing image data from the early days of the reserve to analyze the landscape changes resulting from these effective conservation measures. By 2005 (when the reserve was promoted to a national nature reserve), the proportion of forest landscape in the LNNR had already reached 85.8%, and subsequently, the implementation of various protection projects in the reserve mainly focused on hill closure for forestation. The decline in landscape quality that took place in the LNNR between 2006 and 2019 was mainly due to three reasons. First, there are large industrial and mining enterprises in the reserve, whose activities can cause great damage to the ecological environment [36]. It should be noted that the industrial and mining enterprise existed long before the establishment of the nature reserve (established in 1971), and it was included in the nature reserve when the nature reserve was delimited. The Regulations of the People's Republic of China on Natural Reserves clearly stipulate that industrial and mining activities are prohibited in nature reserves. In 2017, all of the mining enterprises in the LNNR stopped ore mining activities and jointly formulated a "mine vegetation restoration plan" together with the nature reserve management; this plan was implemented by 2018, however, ecological restoration is expected to take some time [37], and the effectiveness of vegetation restoration is not yet clear. Secondly, activities by indigenous residents, such as agricultural reclamation, settlement construction, and resource development have led to a reduction in the quality of the natural ecosystem in the LNNR [38][39][40]. There are large numbers of indigenous residents in the LNNR, and these residents have a relatively concentrated distribution, mainly in the Liancheng town government seat in the southern part of the experimental zone and the Minle Township administrative villages at the eastern edge of the experimental zone. The frequent anthropogenic activities in the LNNR have placed great pressure on the protection and management of the reserve and have become an important reason for the decline of the overall quality of its landscape. Thirdly, the increase in population in the protected area of the LNNR has led to an increase in transport demand, and road construction has inevitably encroached on trees, shrubs, and water, resulting in a deterioration of landscape quality. Most of the PAs in China have a certain degree of human interference because of rapid population growth during the past few decades. However, the current management and protection strategies for PAs are not effective. Therefore, in PAs where habitat quality is declining, diverse ecological restoration projects should be selected, while for PAs that do not have conservation value, management measures should be adjusted in a timely manner to avoid further expansion of the scope and extent of interference.
The PLII can comprehensively reflect the degree of fragmentation, edge effect intensity, and spatial connectivity of wildlife habitats in PAs. This study uses changes in this index to evaluate the effectiveness of the protection of the integrity of the protected landscape in the LNNR. To calculate the index, it is necessary to delineate the protected landscape mosaic according to the real situation in different nature reserves. This is mainly achieved in two ways: First, it is possible to delineate the protected landscape mosaic according to the ecosystem type and protected objects of a specific protected area [41]. For example, in this study, protected landscape types with an ecological level ≥0.38 were combined into a protected landscape mosaic, which included artificial woodland. Meanwhile, related studies showed that in the Wanglang National Nature Reserve, where giant pandas are the main protection object, artificial woodland is not a suitable habitat for the pandas [42]. Second, when calculating the PLII, the interfered landscape which have not caused protected landscape separation can be merged into the protected landscape mosaic for analysis. For example, previous studies considered that transportation land types such as Class IV roads, forest roads, and rural roads do not form a substantial separation to protected landscape patches [41]. Additionally, to ensure the accuracy of the calculation results, the protected landscape mosaic can be further integrated according to the distribution dynamic of protected objects. In summary, the PLII can be used to evaluate the integrity of protected landscapes, and its change value can be employed to assess the effectiveness of PAs for protecting major conservation objects and their habitats.
The ILSI directly reflects the spatial spread of interfered landscape, and this study uses changes in this index to evaluate changes in the effect of anthropogenic interference in the LNNR. For example, in this study, landscape types with an assigned ecological level ≤−0.38 were combined into an interfered landscape mosaic, which includes cultivated land. A previous study showed that in Hanzhong, Shaanxi Province, cultivated land (paddy fields) is an important habitat for the crested ibis (Nipponia nippon) and green rice cultivation is thus of great significance to the conservation of this species [43]. The interfered landscape mosaic can be adjusted according to the main object of protection and its sensitivity to various types of interference, and the ILSI can therefore be universally applied to evaluate and control the spread of human interference.
Based on the results of this study, we give the following recommendations to promote the protection effectiveness of protected areas: (1) adjust the scope of PAs and functional areas as soon as possible, mainly by removing areas without protection value (e.g., towns, industrial and mining enterprises, and water conservancy and hydropower facilities) from PAs to avoid the interference of these areas in the assessment of the effectiveness of the landscape protection of PAs. Furthermore, for interfered landscape that cannot be withdrawn from the scope of PAs, scientific monitoring should be conducted to minimize the negative impact of this landscape on the main protection objects of nature reserves and their habitats; (2) carry out the strict control and environmental impact assessment of the construction of facilities such as roads and hydropower stations in PAs to prevent the increased fragmentation of the protected landscape. At the same time, construct a protected area habitat corridor system to mitigate the decline in the connectivity and integrity of the protected landscape; and (3) obey the National Nature Reserve Regulation of China, follow the management routines of the LNNR strictly, and construct monitoring and response systems for long-term management issues.

Conclusions
Using high-resolution remote sensing image data, this study analyzed the landscape dynamics and its drivers in the LNNR since its promotion to a national-level reserve; specifically, it evaluated the landscape protection effectiveness of the LNNR from 2006 to 2019 using the LTI, PLII, and ILSI. The following conclusions can be drawn: (1) From 2006 to 2019, the area of protected landscapes-namely woodland, grassland, and water-within the protected area of the LNNR decreased from 42,505.59 hm 2 (88.25% of the reserve's total area) to 42,456.16 hm 2 (88.15%), while the area of interfered landscapenamely residential land, industrial and mining land, and water conservancy facility landincreased from 5526.36 hm 2 (11.47% of the reserve's total area) to 5546.71 hm 2 (11.52%), resulting in a decrease in the overall quality of the landscape and an increase in the degree of human interference in the reserve. These changes were mostly concentrated in the experimental zone of the protected area of the LNNR, and the main causes of the overall deterioration of the landscape were industrial and mining exploitation, land use by indigenous inhabitants, and the construction of transportation facilities with contributions of 34.79%, 28.98, and 17.30%, respectively. Although these changes are not obvious, they still require attention and timely remedial measures.
(2) During the same period, there was a decline in the integrity of the protected landscape mosaic in the protected area of the LNNR-the PLII decreased from 82.7 to 68.7. This decline was mainly due to landscape fragmentation caused by road expansion and modification in the experimental zone of the protected area. It is necessary to mitigate the decline in the integrity of the protected landscape of the reserve by constructing ecological corridors and performing ecological restoration.
(3) The ILSI of the LNNR was 26.61 and 26.68 in 2006 and 2019, respectively. The spatial distribution of interfered landscape in the LNNR did not change significantly. The area of interfered landscape increased mostly in areas where such landscape was originally concentrated, and the new patches were mostly small. These results suggest that the protected area of the LNNR is effective for controlling the scope of influence of human interference in the reserve.
This study proposes a comprehensive method for evaluating the landscape conservation effectiveness of PAs based on the LTI, the PLII, and the ILSI. This method has the potential to be applied to the rapid assessment of landscape conservation effectiveness in various types of protected area.
Author Contributions: Z.H. designed the study, data preparation, analysis, and wrote the paper. Y.P. participated in the field investigation and provided suggestions for the method. R.W. participated in the data preparation and processing. B.Z. and N.L. participated in the remote sensing image interpretation and the accuracy calculation of the interpretation results. G.C. provided the project design and performed paper editing. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

PAs
Protected Areas LNNR Liancheng National Nature Reserve LTI Landscape Transfer Index PLII Protected Landscape Integrity Index ILSI Interfered Landscape Sprawl Index